1 The 1-D Heat Equation

Transcription

1 The 1-D Heat Equation Linear Partial Differential Equations Matthew J. Hancock Fall The 1-D Heat Equation 1.1 Physical derivation Reference: Guenther & Lee , Myint-U & Debnath.1 and.5 [Sept. 8, 006] In a metal rod with non-uniform temperature, heat (thermal energy) is transferred from regions of higher temperature to regions of lower temperature. Three physical principles are used here. 1. Heat (or thermal) energy of a body with uniform properties: Heat energy = cmu, where m is the body mass, u is the temperature, c is the specific heat, units [c] = L T U 1 (basic units are M mass, L length, T time, U temperature). c is the energy required to raise a unit mass of the substance 1 unit in temperature.. Fourier s law of heat transfer: rate of heat transfer proportional to negative temperature gradient, Rate of heat transfer u = K 0 (1) area x where K 0 is the thermal conductivity, units [K 0 ] = MLT 3 U 1. In other words, heat is transferred from areas of high temp to low temp. 3. Conservation of energy. Consider a uniform rod of length l with non-uniform temperature lying on the x-axis from x = 0 to x = l. By uniform rod, we mean the density ρ, specific heat c, thermal conductivity K 0, cross-sectional area A are ALL constant. Assume the sides 1

2 of the rod are insulated and only the ends may be exposed. Also assume there is no heat source within the rod. Consider an arbitrary thin slice of the rod of width Δx between x and x+δx. The slice is so thin that the temperature throughout the slice is u (x,t). Thus, Heat energy of segment = c ρaδx u = cρaδxu (x,t). By conservation of energy, change of heat energy of = segment in time Δt heat in from left boundary heat out from right boundary. From Fourier s Law (1), u u cρaδxu (x,t + Δt) cρaδxu (x,t) = ΔtA K 0 ΔtA K 0 x x Rearranging yields (recall ρ, c, A, K 0 are constant), u (x,t + Δt) Δt u (x,t) = K 0 cρ x u x x+δx Δx Taking the limit Δt, Δx 0 gives the Heat Equation, where u t u x x x+δx u = κ () x κ = K 0 (3) cρ is called the thermal diffusivity, units [κ] = L /T. Since the slice was chosen arbitrarily, the Heat Equation () applies throughout the rod. 1. Initial condition and boundary conditions To make use of the Heat Equation, we need more information: 1. Initial Condition (IC): in this case, the initial temperature distribution in the rod u (x, 0).. Boundary Conditions (BC): in this case, the temperature of the rod is affected by what happens at the ends, x = 0,l. What happens to the temperature at the end of the rod must be specified. In reality, the BCs can be complicated. Here we consider three simple cases for the boundary at x = 0.

3 (I) Temperature prescribed at a boundary. For t > 0, u (0,t) = u 1 (t). (II) Insulated boundary. The heat flow can be prescribed at the boundaries, u K 0 (0,t) = φ 1 (t) x (III) Mixed condition: an equation involving u (0,t), u/ x (0,t), etc. Example 1. Consider a rod of length l with insulated sides is given an initial temperature distribution of f (x) degree C, for 0 < x < l. Find u (x,t) at subsequent times t > 0 if end of rod are kept at 0 o C. The Heat Eqn and corresponding IC and BCs are thus PDE: u t = κu xx, 0 < x < l, (4) IC: u (x, 0) = f (x), 0 < x < l, (5) BC: u (0,t) = u (L,t) = 0, t > 0. (6) Physical intuition: we expect u 0 as t. 1.3 Non-dimensionalization Dimensional (or physical) terms in the PDE (): k, l, x, t, u. Others could be introduced in IC and BCs. To make the solution more meaningful and simpler, we group as many physical constants together as possible. Let the characteristic length, time and temperature be L, T and U, respectively, with dimensions [L ] = L, [T ] = T, [U ] = U. Introduce dimensionless variables via x t u (x,t) f (x) xˆ =, tˆ=, u ˆ x, ˆ t ˆ =, fˆ(ˆ x) =. (7) L T U U The variables ˆx, tˆ, û are dimensionless (i.e. no units, [ˆx] = 1). The sensible choice for the characteristic length is L = l, the length of the rod. While x is in the range 0 < x < l, x ˆ is in the range 0 < x ˆ < 1. The choice of dimensionless variables is an ART. Sometimes the statement of the problem gives hints: e.g. the length l of the rod (1 is nicer to deal with than l, an unspecified quantity). Often you have to solve the problem first, look at the solution, and try to simplify the notation. 3

4 From the chain rule, u t = u û tˆ U û = U = t tˆ t T tˆ, u x = u û xˆ U û = U = x xˆ x L xˆ u xx = U û L xˆ Substituting these into the Heat Eqn (4) gives u t = κu xx û T κ û = tˆ L xˆ To make the PDE simpler, we choose T = L /κ = l /κ, so that û û =, 0 < x ˆ < 1, t ˆ> 0. tˆ xˆ The characteristic (diffusive) time scale in the problem is T = l /κ. For different substances, this gives time scale over which diffusion takes place in the problem. The IC (5) and BC (6) must also be non-dimensionalized: IC: û (ˆx, 0) = fˆ(ˆx), 0 < x ˆ < 1, BC: û 0,tˆ = û 1,tˆ = 0, t ˆ> Dimensionless problem Dropping hats, we have the dimensionless problem PDE: u t = u xx, 0 < x < 1, (8) IC: u (x, 0) = f (x), 0 < x < 1, (9) BC: u (0,t) = u (1,t) = 0, t > 0, (10) where x, t are dimensionless scalings of physical position and time. Separation of variables Ref: Guenther & Lee, 4. and 5.1, Myint-U & Debnath 6.4 [Sept 1, 006] We look for a solution to the dimensionless Heat Equation (8) (10) of the form u (x,t) = X (x)t (t) (11) 4

5 Take the relevant partial derivatives: u xx = X (x)t (t), u t = X (x)t (t) where primes denote differentiation of a single-variable function. The PDE (8), u t = u xx, becomes T (t) = T (t) X (x) X (x) The left hand side (l.h.s.) depends only on t and the right hand side (r.h.s.) only depends on x. Hence if t varies and x is held fixed, the r.h.s. is constant, and hence T /T must also be constant, which we set to λ by convention: T (t) X (x) T (t) = X (x) = λ, λ = constant. (1) The BCs become, for t > 0, u (0,t) = X (0)T (t) = 0 u (1,t) = X (1)T (t) = 0 Taking T (t) = 0 would give u = 0 for all time and space (called the trivial solution), from (11), which does not satisfy the IC unless f (x) = 0. If you are lucky and f (x) = 0, then u = 0 is the solution (this has to do with uniqueness of the solution, which we ll come back to). If f (x) is not zero for all 0 < x < 1, then T (t) cannot be zero and hence the above equations are only satisfied if X (0) = X (1) = 0. (13).1 Solving for X (x) Ref: Guenther & Lee, 4. and 5.1 and 7.1, Myint-U & Debnath We obtain a boundary value problem for X (x), from (1) and (13), X (x) + λx (x) = 0, 0 < x < 1, (14) X (0) = X (1) = 0. (15) This is an example of a Sturm-Liouville problem (from your ODEs class). There are 3 cases: λ > 0, λ < 0 and λ = 0. (i) λ < 0. Let λ = k < 0. Then the solution to (14) is X = Ae kx + Be kx 5

6 for integration constants A, B found from imposing the BCs (15), X (0) = A + B = 0, X (1) = Ae k + Be k = 0. The first gives A = B, the second then gives A e k 1 = 0, and since k > 0 we have A = B = u = 0, which is the trivial solution. Thus we discard the case λ < 0. (ii) λ = 0. Then X (x) = Ax +B and the BCs imply 0 = X (0) = B, 0 = X (1) = A, so that A = B = u = 0. We discard this case also. (iii) λ > 0. In this case, (14) is the simple harmonic equation whose solution is X (x) = A cos λx + B sin λx. (16) The BCs imply 0 = X (0) = A, and B sin λ = 0. We don t want B = 0, since that would give the trivial solution u = 0, so we must have sin λ = 0. (17) Thus λ = nπ, for any nonzero integer n (n = 1,, 3,...). We use subscripts to label the particular n-value. The values of λ are called the eigenvalues of the Sturm-Liouville problem (14), λ n = n π, n = 1,, 3,... and the corresponding solutions of (14) are called the eigenfunctions of the Sturm-Liouville problem (14), X n (x) = b n sin (nπx), n = 1,, 3,... (18) We have assumed that n > 0, since n < 0 gives the same solution as n > 0.. Solving for T (t) When solving for X (x), we found that non-trivial solutions arose for λ = n π for all nonzero integers n. The equation for T (t) is thus, from (1), and, for n, the solution is T (t) = n π T (t) where the c n s are constants of integration. n T n = c n e π t, n = 1,, 3,... (19) 6

7 .3 Full solution u (x, t) Ref: Myint-U & Debnath 6.4, Ch 5 Putting things together, we have, from (11), (18) and (19), n u n (x,t) = B n sin (nπx)e π t, n = 1,, 3,... (0) where B n = c n b n. Each function u n (x,t) is a solution to the PDE (8) and the BCs (10). But, in general, they will not individually satisfy the IC (9), u n (x, 0) = B n sin (nπx) = f (x). We now apply the principle of superposition: if u 1 and u are two solutions to the PDE (8) and BC (10), then c 1 u 1 + c u is also a solution, for any constants c 1, c. This relies on the linearity of the PDE and BCs. We will, of course, soon make this more precise... Since each u n (x, 0) is a solution of the PDE, then the principle of superposition says any finite sum is also a solution. To solve the IC, we will probably need all the solutions u n, and form the infinite sum (convergence properties to be checked), u (x,t) = u n (x,t). (1) f (x) = u (x, 0) = u n (x, 0) = B n sin (nπx). () 1 0 m = n 1 sin (mπx) sin (nπx) dx = = δ mn (3) 0 1/ m = n u (x,t) satisfies the BCs (10) since each u n (x,t) does. Assuming term-by-term differentiation holds (to be checked) for the infinite sum, then u (x,t) also satisfies the PDE (8). To satisfy the IC, we need to find B n s such that This is the Fourier Sine Series of f (x). To solve for the B n s, we use the orthogonality property for the eigenfunctions sin (nπx), where δ mn is the kronecker delta, δ mn = 0 m = n 1 m = n The orthogonality relation (3) is derived by substituting sin (mπx) sin (nπx) = cos ((m n) πx) cos ((m + n)πx) 7

8 into the integral on the left hand side of (3) and noting 1 cos (mπx)dx = δ m0. 0 The orthogonality of the functions sin (nπx) is analogous to that of the unit vectors xˆ and ŷ in -space; the integral from 0 to 1 in (3) above is analogous to the dot product in -space. To solve for the B n s, we multiply both sides of () by sin (mπx) and integrate from 0 to 1: 1 1 sin (mπx)f (x)dx = B n sin (nπx) sin (mπx)dx 0 0 Substituting (3) into the right hand side yields 1 1 sin (mπx)f (x)dx = B n δ nm 0 By definition of δ nm, the only term that is non-zero in the infinite sum is the one where n = m, thus 1 1 sin (mπx)f (x) dx = B m 0 Rearranging yields 1 B m = sin (mπx)f (x)dx. (4) 0 The full solution is, from (0) and (1), where B n are given by (4). u (x,t) = n B n sin (nπx)e π t, (5) To derive the solution (5) of the Heat Equation (8) and corresponding BCs (10) and IC (9), we used properties of linear operators and infinite series that need justification. 3 Example : Cooling of a rod from a constant initial temperature Suppose the initial temperature distribution f (x) in the rod is constant, i.e. f (x) = u 0. The solution for the temperature in the rod is (5), n u (x,t) = B n sin (nπx)e π t, 8

9 where, from (4), the Fourier coefficients are given by 1 1 B n = sin (nπx)f (x) dx = u 0 sin (nπx)dx. 0 0 Calculating the integrals gives 1 cos (nπ) 1 B n = u 0 sin (nπx) dx = = u 0 (( 1) n 0 n even u 0 1) = nπ nπ 4u 0 n odd 0 In other words, and the solution becomes 4u 0 B n = 0, B n 1 = (n 1) π u (x,t) = 4u 0 sin ((n 1)πx) exp (n 1) π t. (6) π (n 1) 3.1 Approximate form of solution The number of terms of the series (6) needed to get a good approximation for u (x,t) depends on how close t is to 0. The series is 4u 0 π t sin (3πx) 9π t u (x,t) = sin (πx)e + e +. π 3 The ratio of the first and second terms is second term e 8π t sin 3πx = first term 3 sin πx 8π e t 8 e < using sin nx n sin x 1 for t π It can also be shown that the first term dominates the sum of the rest of the terms, and hence 4u 0 π t 1 u (x,t) sin (πx)e, for t. (7) π π What does t = 1/π correspond to in physical time? In physical time t = l t/κ (recall our scaling - here we use t as the dimensionless time and t as dimensional time), t = 1/π corresponds to: t 15 minutes, for a 1 m rod of copper (κ 1.1 cm sec 1 ) t 169 minutes, for a 1 m rod of steel (κ 0.1 cm sec 1 ) t 47 hours, for a 1 m rod of glass (κ cm sec 1 ) 9 nπ

10 At t = 1/π, the temperature at the center of the rod (x = 1/) is, from (7), 4 u (x,t) u 0 = 0.47u 0. πe Thus, after a (scaled) time t = 1/π, the temperature has decreased by a factor of 0.47 from the initial temperature u Geometrical visualization of the solution [Sept 14, 006] To analyze qualitative features of the solution, we draw various types of curves in D: 1. Spatial temperature profile, given by u = u (x,t 0 ) where t 0 is a fixed value of x. These profiles are curves in the ux-plane.. Temperature profiles in time, u = u (x 0,t) where x 0 is a fixed value of x. These profiles are curves in the ut-plane. 3. Curves of constant temperature in the xt-plane (level curves), u (x,t) = C where C is a constant. Note that the solution u = u (x,t) is a D surface in the 3D uxt-space. The above families of curves are the different cross-sections of this solution surface. Drawing the D cross-sections is much simpler than drawing the 3D solution surface. Sketch typical curves: when sketching the curves in 1-3 above, we draw a few typical curves and any special cases. While math packages such as Matlab can be used to compute the curves from, say, 0 terms in the full power series solution (6), the emphasis in this course is to use simple considerations to get a rough idea of what the solution looks like. For example, one can use the first term approximation (7), simple physical considerations on heat transfer, and the fact that the solution u (x,t) is continuous in x and t, so that if t 1 is close to t 1, u (x,t 1 ) is close to u (x,t ) Spatial temperature profiles For fixed t = t 0, the first term approximate solution (7) is 4u 0 u (x,t) e π t 0 sin (πx), t 1/π. (8) π 10

11 Figure 1: Spatial temperature profiles u(x,t 0 ). This suggests the center of the rod, x = 1/, is a line of symmetry for u (x,t), i.e. 1 1 u + s,t = u s,t and, for each fixed time, the location of the maximum (minimum) temperature if u 0 > 0 (u 0 < 0). We can prove the symmetry property by noting that the original PDE/BC/IC problem is invariant under the transformation x 1 x. Note also that, from (6), u x (x,t) = 4u 0 cos ((n 1)πx) exp (n 1) π t Thus u x (1/,t) = 0 and u xx (1/,t) < 0, so the nd derivative test implies that x = 1/ is a local max. In Figure 1, we have plotted two typical profiles, one at early times t = t 0 0 and the other at late times t = t 0 0, and two special profiles, the initial temperature at t = 0 (u = u 0 ) and the temperature as t (u = 0). The profile for t = t 0 0 is found from the first term approximation (8). The line of symmetry x = 1/ is plotted as a dashed line for reference in Figure 1. 11

12 Figure : Temperature profiles in time u (x 0,t). 3.. Temperature profiles in time Setting x = x 0 in the approximate solution (7), 4u 0 u (x 0,t) sin (πx 0 ) e π t, t 1/π. π Two typical profiles are sketched in Figure, one near the center of the rod (x 0 1/) and one near the edges (x 0 0 or 1). To draw these we noted that the center of the rod cools more slowly than points near the ends. One special profile is plotted, namely the temperature at the rod ends (x = 0, 1) Curves of constant temperature (level curves of u(x,t)) In Figure 3, we have drawn three typical level curves and two special ones, u = 0 (rod ends) and u = u 0 (the initial condition). For a fixed x = x 0, the temperature in the rod decreases as t increases (motivated by the first term approximation (8)), as indicated by the points of intersection on the dashed line. The center of the rod (x = 1/) is a line of symmetry, and at any time, the maximum temperature is at the center. Note that at t = 0, the temperature is discontinuous at x = 0, 1. To draw the level curves, it is easiest to already have drawn the spatial temperature profiles. Draw a few horizontal broken lines across your u vs. x plot. Suppose you draw a horizontal line u = u 1. Suppose this line u = u 1 crosses one of your profiles u (x,t 0 ) at position x = x 1. Then (x 1,t 0 ) is a point on the level curve u (x,t) = u 1. Now plot this point in your level curve plot. By observing where the line u = u 1 crosses your various spatial profiles, you fill in the level curve u (x,t) = u 1. Repeat 1

13 Figure 3: Curves of constant temperature u(x,t) = c, i.e., the level curves of u(x,t). this process for a few values of u 1 to obtain a few representative level curves. Plot also the special cases: u (x,t) = u 0 /w, u = 0, etc. When drawing visualization curves, the following result is also helpful Maximum Principle for the basic Heat Problem Ref: Guenther & Lee 5., Myint-U & Debnath 8. This result is useful when plotting solutions: the extrema of the solution of the heat equation occurs on the space-time boundary, i.e. the maximum of the initial condition and of the time-varying boundary conditions. More precisely, given the heat equation with some initial condition f (x) and BCs u (0,t), u (1,t), then on a given time interval [0,T], the solution u (x,t) is bounded by u min u (x,t) u max where u max = max max f (x), max u (0,t), max u (1,t) 0<x<1 0<t<T 0<t<T u min = min min f (x), min u (0,t), min u (1,t) 0<x<1 0<t<T 0<t<T 13

14 Thus, in the example above, u min = 0 and u max = u 0, hence for all x [0, 1] and t [0,T], 0 u (x,t) u 0. 4 Equilibrium temperature profile (steady-state) Intuition tells us that if the ends of the rod are held at 0 o C and there are no heat sources or sinks in the rod, the temperature in the rod will eventually reach 0. The solution above confirms this. However, we do not have to solve the full problem to determine the asymptotic or long-time behavior of the solution. Instead, the equilibrium or steady-state solution u = u E (x) must be independent of time, and will thus satisfy the PDE and BCs with u t = 0, u E = 0, 0 < x < 1; u E (0) = u E (1) = 0. The solution is u E = c 1 x + c and imposing the BCs implies u E (x) = 0. In other words, regardless of the initial temperature distribution u (x, 0) = f (x) in the rod, the temperature eventually goes to zero. 4.1 Rate of decay of u (x, t) How fast does u approach u E = 0? From our estimate (8) above, u n (x,t) Be n π t, n = 1,, 3,... n where B is a constant. Noting that e n π t e nπ t = e π t, we have n Br π t u n (x,t) u n (x,t) B r =, r = e < 1 (t > 0). 1 r The last step is the geometric series result r n r =, for r < 1. Thus 1 r t Be π u n (x,t), t > 0. (9) 1 e π t Therefore u (x,t) approaches the steady-state u E (x) = 0 exponentially fast (i.e. the rod cools quickly) the first term in the series, sin (πx)e π t (term with smallest eigenvalue λ = π) determines the rate of decay of u (x,t) the B n s may also affect the rate of approach to the steady-state, for other problems 14

15 4. Error of first-term approximation Using the method in the previous section, we can compute the error between the first-term and the full solution (6), u (x,t) u 1 (x,t) = u n (x,t) u 1 (x,t) n Br = u n (x,t) u n (x,t) B r = n= 1 r n= n= where r = e π t. Hence the solution u (x,t) approaches the first term u 1 (x,t) exponentially fast, Be π t u (x,t) u 1 (x,t) 1 e, t > 0. (30) π t With a little more work we can get a much tighter (i.e. better) upper bound. We consider the sum Substituting for u n (x,t) gives u n (x,t) n=n n u n (x,t) B e π t = Be N π t n=n n=n n=n Here s the trick: for n N, (n e N )π t (31) n N = (n + N) (n N) N (n N) 0. Since e x is a decreasing function, then (n e N )π t N(n N)π e t (3) for t > 0. Using the inequality (3) in Eq. (31) gives u n (x,t) Be N π t n=n n=n N(n N)π e t N Nπ t n e = Be N π t Nπ e t = Be N π t. 1 e Nπ t n=n Simplifying the expression on the right hand side yields n=n Be N π t u n (x,t) 1 e Nπ t (33) 15

16 For N = 1, (33) gives the temperature decay t Be π u (x,t) = u n (x,t) u n (x,t), (34) 1 e π t n=n which is slightly tighter than (9). For N =, (33) gives the error between u (x,t) and the first term approximation, Be 4π t u (x,t) u 1 (x,t) = u n (x,t) u n (x,t) 1 e 4π t n=n n= n= which is a much better upper bound than (30). Note that for an initial condition u (x, 0) = f (x) that is is symmetric with respect to x = 1/ (e.g. f (x) = u 0 ), then u n = 0 for all n, and hence error between u (x,t) and the first term u 1 (x,t) is even smaller, u (x,t) u 1 (x,t) = u n (x,t) u n (x,t) Applying result (33) with N = 3 gives n=3 n=3 Be 9π t u (x,t) u 1 (x,t). 1 e 6π t Thus, in this case, the error between the solution u (x,t) and the first term u 1 (x,t) decays as e 9π t - very quickly! 5 Review of Fourier Series Ref: Guenther & Lee, 3.1, Myint-U & Debnath , [Sept 19, 006] Motivation: Recall that the initial temperature distribution satisfies f (x) = u (x, 0) = B n sin (nπx). In the example above with a constant initial temperature distribution, f (x) = u 0, we have 4u 0 sin ((n 1) πx) u 0 =. (35) π n 1 Note that at x = 0 and x = 1, the r.h.s. does NOT converge to u 0 = 0, but rather to 0 (the BCs). Note that the Fourier Sine Series of f (x) is odd and -periodic in space and converges to the odd periodic extension of f (x) = u 0. 16

17 Odd periodic extension The odd periodic extension of a function f (x) defined for x [0, 1], is the Fourier Sine Series of f (x) evaluated at any x R, f (x) = B n sin (nπx). Note that since sin (nπx) = sin ( nπx) and sin (nπx) = sin (nπ (x + )), then f (x) = f ( x) and f (x + ) = f (x). Thus f (x) is odd, -periodic and f (x) equals f (x) on the open interval (0, 1). What conditions are necessary for f (x) to equal f (x) on the closed interval [0, 1]? This is covered next. Aside: cancelling u 0 from both sides of (35) gives a really complicated way of writing 1, 5.1 Fourier Sine Series 4 sin ((n 1)πx) 1 =. π n 1 Given an integrable function f (x) on [0, 1], the Fourier Sine Series of f (x) is B n sin (nπx) (36) where 1 B n = f (x) sin (nπx)dx. (37) 0 The associated orthogonality properties are 1 1/, m = n = 0, sin (mπx) sin (nπx)dx = 0 0, m = n. 5. Fourier Cosine Series Given an integrable function f (x) on [0, 1], the Fourier Cosine Series of f (x) is where A 0 + A n cos (nπx) (38) 1 A 0 = f (x)dx, (average of f (x)) (39) 0 1 A n = f (x) cos (nπx) dx, n 1. (40) 0 17

18 The associated orthogonality properties are 1 1/, m = n = 0, cos (mπx) cos (nπx) dx = 0 0, m = n. e.g. The cosine series of f (x) = u 0 for x [0, 1] is just u 0. In other words, A 0 = u 0 and A n = 0 for n The (full) Fourier Series The Full Fourier Series of an integrable function f (x), now defined on [ 1, 1], is where fˆ(x) = a 0 + (a n cos (nπx) + b n sin (nπx)) (41) 1 a 0 = 1 f (x) dx 1 1 a n = f (x) cos (nπx)dx 1 1 b n = f (x) sin (nπx)dx 1 The associated orthogonality properties of sin and cos on [ 1, 1] are, for any m,n = 1,, 3,... 1 sin (mπx) cos (nπx)dx = 0, Piecewise Smooth Ref: Guenther & Lee p. 50, all m,n, 1 1, m = n = 0, sin (mπx) sin (nπx)dx = 1 0, m = n, 1 1, m = n = 0, cos (mπx) cos (nπx) dx = 1 0, m = n. Provided a function f (x) is integrable, its Fourier coefficients can be calculated. It does not follow, however, that the corresponding Fourier Series (Sine, Cosine or Full) converges or has the sum f (x). In order to ensure this, f (x) must satisfy some stronger conditions. 18

19 Definition Piecewise Smooth function: A function f (x) defined on a closed interval [a,b] is said to be piecewise smooth on [a,b] if there is a partition of [a,b], a = x 0 < x 1 < x < < x n = b such that f has a continuous derivative (i.e. C 1 ) on each closed subinterval [x m,x m+1 ]. [a,b]. E.g. any function that is C 1 on an interval [a,b] is, of course, piecewise smooth on E.g. the function f (x) = x, 0 x 1/ 1/, 1/ < x 1 is piecewise smooth on [0, 1], but is not continuous on [0, 1]. E.g. the function f (x) = x is both continuous and piecewise smooth on [ 1, 1], despite f (x) not being defined at x = 0. This is because we partition [ 1, 1] into two subintervals [ 1, 0] and [0, 1]. When worrying about f (x) near x = 0, note that on [0, 1] we only care about the left limit f (0 ) and for [ 1, 0], we only care about the right limit f (0 + ). E.g. the function f (x) = x 1/ on [ 1, 1], since f (0 ) and f (0 + ) do not exist. is continuous on [ 1, 1] but not piecewise smooth 5.5 Convergence of Fourier Series Ref: Guenther & Lee p. 49 and (optional) 3.3, Myint-U & Debnath 5.10 Theorem [Convergence of the Fourier Sine and Cosine Series]: If f (x) is piecewise smooth on the closed interval [0, 1] and continuous on the open interval (0, 1), then the Fourier Sine and Cosine Series converge for all x [0, 1] and have the sum f (x) for all x (0, 1). Note: Suppose f (x) is piecewise smooth on [0, 1] and is continuous on (0, 1) except at a jump discontinuity at x = a. Then the Fourier Sine and Cosine Series converge to f (x) on (0, 1) and converge to the average of the left and right limits at x = a, i.e. (f (a ) + f (a+))/. At the endpoints x = 0, 1, the Sine series converges to zero, since sin (nπx) = 0 at x = 0, 1 for all n. Theorem [Convergence of the Full Fourier Series]: If f (x) is piecewise smooth on the closed interval [ 1, 1] and continuous on the open interval ( 1, 1), then the Full Fourier Series converges for all x [ 1, 1] and has the sum f (x) for all x ( 1, 1). 5.6 Comments Ref: see problems Guenther & Lee p

20 Given a function f (x) that is piecewise smooth on [0, 1] and continuous on (0, 1), the Fourier Sine and Cosine Series of f (x) converge on [0, 1] and equal f (x) on the open interval (0, 1) (i.e. perhaps excluding the endpoints). Thus, for any x (0, 1), A 0 + A n cos (nπx) = f (x) = B n sin (nπx). In other words, the Fourier Cosine Series (left hand side) and the Fourier Sine Series (right hand side) are two different representations for the same function f (x), on the open interval (0, 1). The values at the endpoints x = 0, 1 may not be the same. The choice of Sine or Cosine series is determined from the type of eigenfunctions that give solutions to the Heat Equation and BCs. 6 Well-Posed Problems Ref: Guenther & Lee 1.8 and 5. (in particular p. 160 and preceding pages, and p. 150) Ref: Myint-U & Debnath 1., 6.5 We call a mathematical model or equation or problem well-posed if it satisfies the following 3 conditions: 1. [Existence of a solution]: The mathematical model has at least 1 solution. Physical interpretation: the system exists over at least some finite time interval.. [Uniqueness of solution]: The mathematical model has at most 1 solution. Physical interpretation: identical initial states of the system lead to the same outcome. 3. [Continuous dependence on parameters]: The solution of the mathematical model depends continuously on initial conditions and parameters. Physical interpretation: small changes in initial states (or parameters) of the system produce small changes in the outcome. If an IVP (initial value problem) or BIVP (boundary initial value problem - e.g. Heat Problem) satisfies 1,, 3 then it is well-posed. Example: for the basic Heat Problem, we showed 1 by construction a solution using the method of separation of variables. Continuous dependence is more difficult to show (need to know about norms), but it is true, and we will use this fact when sketching solutions. Also, when drawing level curves u (x,t) = const, small changes in parameters (x,t) leads to a small change in u. We now prove the nd part of well-posedness, uniqueness of solution, for the basic heat problem. 0

21 6.1 Uniqueness of solution to the Heat Problem Definition We define the two space-time sets and the space of functions D = {(x,t) : 0 x 1, t > 0} D = {(x,t) : 0 x 1, t 0} C D = u (x,t) : uxx continuous in D and u continuous in D. In other words, the space of functions that are twice continuously differentiable on [0, 1] for t > 0 and continuous on [0, 1] for t 0. Theorem The basic Heat Problem, i.e. the Heat Equation (8) with BC (10) and IC (9), PDE: u t = u xx, 0 < x < 1, IC: u (x, 0) = f (x), 0 < x < 1, BC: u (0,t) = u (1,t) = 0, t > 0, has at most one solution in the space of functions C D. Proof: Consider two solutions u 1, u C D to the Heat Problem. Let v = u 1 u. We aim to show that v = 0 on [0, 1], which would prove that u 1 = u and then solution to the Heat Equation (8) with BC (10) and IC (9) is unique. Since each of u 1, u satisfies (8), (9), and (10), the function v satisfies and similarly, v t = (u 1 u ) (4) t = u 1t u t = u 1xx u xx = (u 1 u ) xx = v xx, 0 < x < 1, IC: v (x, 0) = u 1 (x, 0) u (x, 0) = f (x) f (x) = 0, 0 < x < 1, (43) BC: v (0,t) = u 1 (0,t) u (0,t) = 0, v (1,t) = 0, t > 0. (44) Define the function V (t) = 1 0 v (x,t)dx 0, t 0. 1

22 Showing that v (x,t) = 0 reduces to showing that V (t) = 0, since v (x,t) is continuous on [0, 1] for all t 0 and if there was a point x such that v (x,t) = and substitute for v t from the PDE (4), 1 1 dv = vv t dx = vv xx dx Integrating by parts (note: vv xx = (vv x ) x v x ) gives dv 1 1 = vv xx dx = vv x dt x=0 1 dv = v dt x dx 0. v x dx 0, then V (t) would be strictly greater than 0. To show V (t) = 0, we differentiate V (t) in time Using the BCs (44) gives The IC (43) implies that dt V (0) = v (x, 0) dx = 0. 1 Thus, V (t) V /dt V (0) = 0, i.e. 0, d 0, and V (t) is a non-negative, nonincreasing function of time whose initial value is zero. Thus, for all time, V (t) = 0 and v (x,t) = 0 for all x [0, 1], implying that u 1 = u. This proves that the solution to the Heat Equation (8), its IC (9), and BCs (10) is unique, i.e. there is at most one solution. Uniqueness proofs for other types of BCs follows in a similar manner. 0 7 Variations on the basic Heat Problem [Sept 1, 006] We now consider variations to the basic Heat Problem, including different types of boundary conditions and the presence of sources and sinks. 7.1 Boundary conditions Type I BCs (Dirichlet conditions) Ref: Guenther & Lee p. 149 Type I, or Dirichlet, BCs specify the temperature u (x,t) at the end points of the rod, for t > 0, u (0,t) = g 1 (t), u (1,t) = g (t).

23 Type I Homogeneous BCs are u (0,t) = 0, u (1,t) = 0. The physical significance of these BCs for the rod is that the ends are kept at 0 o C. The solution to the Heat Equation with Type I BCs was considered in class. After separation of variables u (x,t) = X (x) T (t), the associated Sturm-Liouville Boundary Value Problem for X (x) is X + λx = 0; X (0) = X (1) = 0. The eigenfunctions are X n (x) = B n sin (nπx) Type II BCs (Newmann conditions) Ref: Guenther & Lee p. 15 problem 1 Type II, or Newmann, BCs specify the rate of change of temperature u/ x (or heat flux) at the ends of the rod, for t > 0, u (0,t) = g 1 (t), x u (1,t) = g (t). x Type II Homogeneous BCs are u (0, t) = 0, x u (1,t) = 0. x The physical significance of these BCs for the rod is that the ends are insulated. These lead to another relatively simple solution involving a cosine series (see problem 6 on PS 1). After separation of variables u (x,t) = X (x)t (t), the associated Sturm- Liouville Boundary Value Problem for X (x) is X + λx = 0; X (0) = X (1) = 0. The eigenfunctions are X 0 (x) = A 0 = const and X n (x) = A n cos (nπx) Type III BCs (Mixed) The general Type III BCs are a mixture of Type I and II, for t > 0, u α 1 (0,t) + α u (0,t) = g 1 (t), x u α 3 (1,t) + α 4 u (1,t) = g (t). x 3

24 After separation of variables u (x,t) = X (x)t (t), the associated Sturm-Liouville Boundary Value Problem for X (x) is X + λx = 0, α 1 X (0) + α X (0) = 0 α 3 X (0) + α 4 X (0) = 0 The associated eigenfunctions depend on the values of the constants α 1,,3,4. Example 1. α 1 = α 4 = 1, α = α 3 = 0. Then X + λx = 0; X (0) = X (1) = 0 n and the eigenfunctions are X n = A n cos 1 πx. Note: the constant X 0 = A 0 is not an eigenfunction here. Example. α 1 = α 4 = 0, α = α 3 = 1. Then X + λx = 0; X (0) = X (1) = 0 n and the eigenfunctions are X n = B n sin 1 πx. Note: the constant X 0 = A 0 is not an eigenfunction here. Note: Starting with the BCs in Example 1 and rotating the rod about x = 1/, you d get the BCs in Example. It is not surprising then that under the change of variables x = 1 x, Example 1 becomes Example, and vice versa. The eigenfunctions also possess this symmetry, since n 1 sin π (1 x) = ( 1) n n 1 cos πx. Since we can absorb the ( 1) n into the constant B n, the eigenfunctions of Example 1 become those of Example under the transformation x = 1 x, and vice versa. 7. Solving the Heat Problem with Inhomogeneous (timeindependent) BCs Ref: Guenther & Lee p. 149 Consider the Heat Problem with inhomogeneous Type I BCs, u t = u xx, 0 < x < 1 u (0,t) = 0, u (1,t) = u 1 = const, t > 0, (45) u (x, 0) = 0, 0 < x < 1. 4

25 Directly applying separation of variables u (x,t) = X (x)t (t) is not useful, because we d obtain X (1)T (t) = u 1 for t > 0. The strategy is to rewrite the solution u (x,t) in terms of a new variable v (x,t) such that the new problem for v has homogeneous BCs! Step 1. Find the steady-state, or equilibrium solution u E (x), since this by definition must satisfy the PDE and the BCs, u E = 0, 0 < x < 1 u E (0) = 0, Solving for u E gives u E (x) = u 1 x. u E (1) = u 1 = const. Step. Transform variables by introducing a new variable v (x,t), v (x,t) = u (x,t) u E (x) = u (x,t) u 1 x. (46) Substituting this into the Heat Problem (45) gives a new Heat Problem, v t = v xx, 0 < x < 1 v (0,t) = 0, v (1,t) = 0, t > 0, (47) v (x, 0) = u 1 x, 0 < x < 1. Notice that the BCs are now homogeneous, and the IC is now inhomogeneous. Notice also that we know how to solve this - since it s the basic Heat Problem! Based on our work, we know that the solution to (47) is 1 n v (x,t) = B n sin (nπx)e π t, B n = f (x) sin (nπx) dx (48) 0 where f (x) = u 1 x. Substituting for f (x) and integrating by parts, we find 1 B n = u 1 x sin (nπx)dx x cos (nπx) 1 = u 1 nπ + cos (nπx)dx x=0 nπ 0 u 1 ( 1) n = (49) nπ Step 3. Transform back to u (x,t), from (46), (48) and (49), u (x,t) = u E (x) + v (x,t) = u 1 x + u 1 π ( 1) n sin (nπx)e n n π t. (50) The term u E (x) is the steady state and the term v (x,t) is called the transient, since it exists initially to satisfy the initial condition but vanishes as t. You can check for yourself by direct substitution that Eq. (50) is the solution to the inhomogeneous Heat Problem (45), i.e. the PDE, BCs and IC. 5

26 7.3 Heat sources Derivation Ref: Guenther & Lee p. 6 (Eq. 3-3) To add a heat source to the derivation of the Heat Equation, we modify the energy balance equation to read, change of heat energy of = segment in time Δt heat in from heat out from heat generated + left boundary right boundary in segment. Let Q (x,t) be the heat generated per unit time per unit volume at position x in the rod. Then the last term in the energy balance equation is just QAΔxΔt. Applying Fourier s Law (1) gives u u cρaδxu (x,t + Δt) cρaδxu (x,t) = ΔtA K 0 ΔtA K 0 x x +QAΔxΔt AΔxΔt and rearranging yields u u u (x,t + Δt) u (x,t) K 0 x x+δx x x Q = +. Δt cρ Δx cρ x x+δx The last term is new; the others we had for the rod without sources. Dividing by Taking the limit Δt, Δx 0 gives the Heat Equation with a heat source, u u Q = κ + (51) t x cρ Introducing non-dimensional variables x = x/l, t = κt/l gives u u l Q = + (5) t x κcρ Defining the dimensionless source term q = l Q/ (κcρ) and dropping tildes gives the dimensionless Heat Problem with a source, u u = + q (53) t x 6

27 7.3. Solution method Ref: Guenther & Lee p , Myint-U & Debnath 6.7 (exercises) The simplest case is that of a constant source q = q (x) in the rod. The Heat Problem becomes, u t = u xx + q (x), 0 < x < 1, (54) u (0,t) = b 1, u (1,t) = b, t > 0, u (x, 0) = f (x), 0 < x < 1. If q (x) > 0, heat is generated at x in the rod; if q (x) < 0, heat is absorbed. The solution method is the same as that for inhomogeneous BCs: find the equilibrium solution u E (x) that satisfies the PDE and the BCs, 0 = u E + q (x), 0 < x < 1, u E (0) = b 1, u E (1) = b. Then let v (x,t) = u (x,t) u E (x). Substituting u (x,t) = v (x,t) + u E (x) into (54) gives a problem for v (x,t), v t = v xx, 0 < x < 1, v (0,t) = 0, v (1,t) = 0, t > 0, v (x, 0) = f (x) u E (x), 0 < x < 1. Note that v (x,t) satisfies the homogeneous Heat Equation (PDE) and homogeneous BCs, i.e. the basic Heat Problem. Solve the Heat Problem for v (x,t) and then obtain u (x,t) = v (x,t) + u E (x). Note that things get complicated if the source is time-dependent - we won t see that in this course. 7.4 Periodic boundary conditions Ref: Guenther & Lee p ; for alternate method, see Guether & Lee p. 149 and then p. 147 [Sept 6, 006] Above, we solved the heat problem with inhomogeneous, but time-independent, BCs by using the steady-state. We now show how to solve the heat problem with inhomogeneous, but time-varying, BCs. We consider the heat problem with an oscillatory 7

28 (and periodic) BC, u t = u xx, 0 < x < 1 u (0,t) = A cos t, u (1,t) = 0, t > 0, (55) u (x, 0) = f (x), 0 < x < 1. The physical meaning of the BC u (0,t) = A cos t is that we keep changing, in a periodic fashion, the temperature at the end x = 0 of the rod. We don t expect the solution to be independent of time as t, since we re changing the temperature periodically at one end. However, we do expect that after an initial transient time, the solution will become periodic with angular frequency, i.e. u (x,t) = v (x,t) + A (x) cos (t + φ (x)) where v (x,t) 0 as t is the transient, A (x) cos (t + φ (x)) is what we call the quasi-steady state, A (x) and φ (x) are the amplitude and phase of the quasi steady state. To solve the problem, the goal is to first find A (x) and φ (x), and then v (x,t), if necessary. Often we might not care about the transient state, if we are more interested in the solution after long times Complexify the problem We use the notation Re {z}, Im {z} to denote the real and imaginary parts of a complex number z. Note that 1 iθ Re {z} = (z + z ), cos θ = Re e 1 Im {z} = (z z ), sin θ = Im e iθ i where asterisks denote the complex conjugate ((x + iy) = x iy). Thus, we can write our quasi-steady solution in terms of complex exponentials, A (x) cos (t + φ (x)) = Re A (x)eiφ(x) it it e = Re U (x)e where, for convenience, we have replaced A (x)e iφ(x) with the complex function U (x). We do this because complex exponentials are much easier to work with than cos (t) and sin (t). Note that U (x) has magnitude A (x) = U (x) and phase φ (x) = Im{U(x)} arctan. The phase φ (x) delays the effects of what is happening at the end Re{U(x)} of the rod: if the end is heated at time t = t 1, the effect is not felt at the center until a later time t = φ (1/) / + t 1. The following result will be useful. 8

29 Lemma [Zero sum of complex exponentials] If, for two complex constants a, b, we have for all times t in some open interval, then a = b = 0. Proof: Differentiate (56) in time t, it ae + be it = 0 (56) i ae it be it = 0. (57) Adding (56) to 1/i (57) gives it ae = 0. Since e it is never zero ( e it = 1), then a = 0. From (56), be it = 0 and hence b = 0. Note that we could also use the Wronskian to show this: it it W e it e e,e it = det = i = 0 ie it ie it and hence e it and e it are linearly independent, meaning that a = b = ODE and ICs for quasi-steady state Step 1. Find the quasi-steady state solution to the PDE and BCs of the Heat Problem (55) of the form u SS (x,t) = Re U (x)e it 1 = U (x)e it + U (x)e it = A (x) cos (t + φ (x)) (58) where U (x) is a complex valued function. Substituting (58) for u (x,t) into the PDE in (55) gives 1 it it iu (x) e iu (x)e = 1 it it U (x)e + U (x)e, for 0 < x < 1 and t > 0. Multiplying both sides by and re-grouping terms yields it it (iu (x) U (x))e + ( iu (x) U (x))e = 0, 0 < x < 1, t > 0. Applying the Lemma to (59) gives (59) iu (x) U (x) = 0 = iu (x) U (x) (60) 9

30 Note that the left and right hand sides are the complex conjugates of one another, and hence they both say the same thing (so from now on we ll write one or the other). Substituting (58) into the BCs in (55) gives 1 A 1 it it it it it it U (0)e + U (0)e = e + e, U (1)e + U (1)e = 0, (61) for t > 0. Grouping the coefficients of e ±it and applying the Lemma yields U (0) = A, U (1) = 0. (6) To summarize, the problem for the complex amplitude U (x) of the quasi-steadystate u SS (x,t) is, from (60) and (6), U (x) iu (x) = 0; U (0) = A, U (1) = 0. (63) Note that (1 + i) = i, and hence Therefore, (63) can be rewritten as 1 i = (1 + i) = (1 + i). U (1 + i) U = 0; U (0) = A, U (1) = 0. (64) Solving for quasi-steady state Solving the ODE (64) gives U = c 1 exp (1 + i)x + c exp (1 + i)x (65) where c 1, c are integration constants. Imposing the BCs gives A = U (0) = c 1 +c, 0 = U (1) = c 1 exp (1 + i) +c exp (1 + i). Solving this set of linear equations for the unknowns (c 1,c ) and substituting these back into gives A exp (1 + i) c 1 = exp (1 + i) exp (1 + i) A exp (1 + i) c = A c 1 = exp (1 + i) exp (1 + i) 30

31 1 0 U(x) /A phase(u(x)) x x Figure 4: At left, the magnitude of U(x) (solid) and U(x) (dash). At right, the phase of U(x). Substituting these into (65) gives exp (1 + i) (1 x) (1 + i) (1 x) exp U = A. exp (1 + i) exp (1 + i) Therefore, the quasi-steady-state solution to the heat problem is exp (1 + i) (1 x) (1 + i) (1 x) u exp SS (x,t) = Re exp (1 + i) exp (1 + i) Ae it It is easy to check that u SS (x,t) satisfies the PDE and BCs in (55). Also, note that the square of the magnitude of the denominator is (1 + i) exp exp (1 + i) = cosh cos > 0 which is greater than zero since > 0 and hence cosh > 1 cos. In Figure 4, the magnitude and phase of U (x) are plotted as solid lines. The straight dashed line is drawn with U (x) for comparison, illustrating that U (x) is nearly linear in x. The phase of U (x) is negative, indicating a delay between what happens at a point x on the rod and what happens at the end x = Solving for the transient Step. Solve for the transient, defined as before, v (x,t) = u (x,t) u SS (x,t). (66) 31

32 Substituting (66) into the heat problem (55), given that u SS (x,t) satisfies the PDE and BCs in (55), gives the following problem for v (x,t), v t = v xx, 0 < x < 1 v (0,t) = 0, v (1,t) = 0, t > 0, (67) v (x, 0) = f (x), 0 < x < 1, where the initial condition f (x) is given by f (x) = u (x, 0) u SS (x, 0) (1 + i) (1 x) exp (1 + i) (1 x) = f (x) Re A exp. exp (1 + i) exp (1 + i) The problem for v (x,t) is the familiar basic Heat Problem whose solution is given by (5), (4) with f (x) replaced by f (x), v (x,t) = 1 n π t B n sin (nπx) e Full solution The full solution to the problem is exp (1 + i) (1 x) exp (1 + i) (1 x) u (x,t) = Re Ae it exp (1 + i) exp (1 + i) + B n sin (nπx)e n π t, B n = f (x) sin (nπx)dx.. The first term is the quasi-steady state, whose amplitude at each x is constant, plus a transient part v (x,t) that decays exponentially as t. If the IC f (x) were given, then we could compute the B n s Similar problem: heating/cooling of earth s surface Consider a vertical column in the earth s crust that is cooled in the winter and heated in the summer, at the surface. We take the x-coordinate to be pointing vertically downward with x = 0 corresponding to the earth s surface. For simplicity, we model the column of earth by the semi-infinite line 0 x <. We crudely model the heating and cooling at the surface as u (0,t) = A cos t, where = π/τ and the (scaled) period τ corresponds to 1 year. Under our scaling, τ = κ (1 year)/l. The boundary condition as x is that the temperature u is bounded ( is at the 3

33 bottom of the earth s crust, still far away from the core, whose effects are neglected). What is the quasi-steady state? The quasi-steady state satisfies the Heat Equation and the BCs, (u SS ) t = (u SS ) xx, 0 < x < (68) u SS (0,t) = T 0 + T 1 cos t, u SS bounded as x, t > 0. We use superposition: u (x,t) = u 0 + u 1, where (u 0 ) t = (u 0 ) xx, (u 1 ) t = (u 1 ) xx, 0 < x <, t > 0. u 0 (0,t) = T 0, u 1 (0,t) = T 1 cos t, u 0, u 1 bounded as x, Obviously, u 0 (x,t) = T 0 works, and by uniqueness, we know this is the only solution for u 0 (x,t). To solve for u 1, we proceed as before and let u 1 (x,t) = Re {U (x) e it } to obtain U (x) iu (x) = 0, 0 < x < (69) The general solution to the ODE (69) is U (0) = T 1, U bonded as x, t > 0. U = c 1 exp (1 + i)x + c exp (1 + i)x. The boundedness criterion gives c = 0, since that term blows up BC at the surface (x = 0) gives c 1 = T 1. Hence U = T 1 exp (1 + i)x. as x. The Putting things together, we have u SS (x,t) = T 0 + Re T 1 exp (1 + i)x e it x = T 0 + T 1 e Re exp i x + it x = T 0 + T 1 e cos x + t. (70) u SS (x,t) is plotted at various dimensionless times t/π = 0, 1/4, 1/, 3/4, 1 in Figure 5. Dashed lines give the amplitude T x 0 ±T 1 e of the quasi-steady-state u SS (x,t). Physical questions: What is the ideal depth for a wine cellar? We want the wine to be relatively cool compared to the summer temperature and relatively warm to the 33

34 x u SS (x,t) Figure 5: Plot of u SS (x,t) at various times. Numbers in figure indicate t/π. 34

35 winter temperature, and yet we want the cellar close to the surface (to avoid climbing too many stairs). Then we must find the smallest depth x such that the temperature u SS (x,t) will be opposite in phase to the surface temperature u SS (0,t). We take κ = 10 3 cm /s and l = 1 m. Recall that the period is 1 year, τ = (κ/l ) (1 year) and 1 year is s. From the solution (70), the phase of u SS (x,t) is reversed when Solving for x gives x = π x = π / Returning to dimensional coordinates, we have x = lx = lπ τ = πκ (1 year) = π ( 10 3 cm ) = 4.45 m. π At this depth, the amplitude of temperature variation is x π T 1 e = T 1 e 0.04T 1 Thus, the temperature variations are only 4% of what they are at the surface. And being out of phase with the surface temperature, the temperature at x = 4.45 m is cold in the summer and warm in the winter. This is the ideal depth of a wine cellar. Note: for a different solution to this problem using Laplace transforms (not covered in this course), see Myint-U & Debnath Example Linearity, Homogeneity, and Superposition Ref: Myint-U & Debnath 1.1, 1.3, 1.4 [Sept 8, 006] Definition Linear space: A set V is a linear space if, for any two elements 1 v 1, v V and any scalar (i.e. number) k R, the terms v 1 +v and kv 1 are also elements of V. E.g. Let S denote the set of functions that are C (twice-continuously differentiable) in x and C 1 (continuously differentiable) in t, for x [0, 1], t 0. We write S as S = {f (x,t) f xx, f t continuous for x [0, 1], t 0} (71) If f 1,f S, i.e. f 1,f are functions that have continuous second derivatives in space and continuous derivatives in time, and k R, then f 1 + f and kf 1 also have the 1 Note that the symbol means an element of. So x [0,l] means x is in the interval [0,l]. 35

36 same, and hence are both elements of S. Thus, by definition, S forms a linear space over the real numbers R. For instance, f 1 (x,t) = sin (πx)e π t and f = x + t 3. Definition Linear operator: An operator L : V W is linear if L (v 1 + v ) = L (v 1 ) + L (v ), (7) L (kv 1 ) = kl (v 1 ), (73) for all v 1,v V, k R. The first property is the summation property; the second is the scalar multiplication property. The term operator is very general. It could be a linear transformation of vectors in a vector space, or the derivative operation / x acting on functions. Differential operators are just operators that contain derivatives. So / x and / x + / y are differential operators. These operators are functions that act on other functions. For example, the x-derivative operation is linear on a space of functions, even though the functions the derivative acts on might not be linear in x: cos x + x = (cos x) + x. x x x E.g. the identity operator I, which maps each element to itself, i.e. for all elements v V we have I (v) = v. Check for yourself that I satisfies (7) and (73) for an arbitrary linear space V. E.g. consider the partial derivative operator acting on S (defined in (71)), u D (u) =, x u S. From the well-known properties of the derivative, namely u v (u + v) = +, u,v S, x x x u (ku) = k, u S,k R x x it follows that the operator D satisfies properties (7) and (73) and is therefore a linear operator. E.g. Consider the operator that defines the Heat Equation L = t x and acts over the linear space S of functions C in x, C 1 in t. The Heat Equation can be written as u u L (u) = = 0. t x 36

37 For any two functions u,v S and a real k R, L (u + v) = (u + v) t x u u v v = (u + v) (u + v) = + = L (u) + L (v) t x t x t x u u L (ku) = (ku) = (ku) (ku) = k = kl (u) t x t x t x Thus, L satisfies properties (7) and (73) and is therefore a linear operator. 8.1 Linear and homogeneous PDE, BC, IC Consider a differential operator L and operators B 1, B, I that define the following problem: PDE: L (u) = h (x,t), with BC: B 1 (u (0,t)) = g 1 (t), B (u (1,t)) = g (t) IC: I (u (x, 0)) = f (x). Definition The PDE is linear if the operator L is linear. Definition The PDE is homogeneous if h (x,t) = 0. Definition The BCs are linear if B 1, B are linear. The BCs are homogeneous if g 1 (t) = g (t) = 0. Definition The IC is linear if I is a linear. The IC is homogeneous if f (x) = The Principle of Superposition The Principle of Superposition has three parts, all which follow from the definition of a linear operator: (1) If L (u) = 0 is a linear, homogeneous PDE and u 1, u are solutions, then c 1 u 1 + c u is also a solution, for all c 1, c R. () If u 1 is a solution to L (u 1 ) = f 1, u is a solution to L (u ) = f, and L is linear, then c 1 u 1 + c u is a solution to L (u) = c 1 f 1 + c f, for all c 1, c R. (3) If a PDE, its BCs, and its IC are all linear, then solutions can be superposed in a manner similar to (1) and (). 37

38 (1) is The proofs of (1)-(3) follow from the definitions above. For example, the proof of L (c 1 u 1 + c u ) = L (c 1 u 1 ) + L (c u ) = c 1 L (u 1 ) + c L (u ) = c c 0 = 0 The first step follows from the addition rule of linear operators; the second from the scalar multiplication rule, and the third from the fact that u 1 and u are solutions of L (u) = 0. The proofs of parts () and (3) are similar. E.g. Suppose u 1 is a solution to and u is a solution to Then u 1 u would solve L (u) = u t u xx = 0 B 1 (u (0,t)) = u (0,t) = 0 B (u (1,t)) = u (1,t) = 0 I (u (x, 0)) = u (x, 0) = 100 L (u) = u t u xx = 0 B 1 (u (0,t)) = u (0,t) = 100 B (u (1,t)) = u (1,t) = 0 I (u (x, 0)) = u (x, 0) = 0 L (u) = u t u xx = 0 B 1 (u (0,t)) = u (0,t) = 100 B (u (1,t)) = u (1,t) = 0 I (u (x, 0)) = u (x, 0) = 00 To solve L (u) = g with inhomogeneous BCs using superposition, we can solve two simpler problems: L (u) = g with homogeneous BCs and L (u) = 0 with inhomogeneous BCs. 8.3 Application to the solution of the Heat Equation Recall that we showed that u n (x,t) satisfied the PDE and BCs. Since the PDE (8) and BCs (10) are linear and homogeneous, then we apply the Principle of Superposition (1) repeatedly to find that the infinite sum u (x,t) given in (80) is also a solution, provided of course that it can be differentiated term-by-term. 38